Abstract: A multi-faceted simulation research study was conducted, addressing several waterways in the ports of Hampton Roads, Virginia. The research, which supported the Norfolk Harbor 50' deepening project, addressed the following elements: Thimble Shoal Channel, Atlantic Ocean Channel, and Norfolk Harbor Reach. The study approach was based on consideration of the overall project requirements together with those of the individual elements, resulting in a research methodology tailored for each element. Methodological characteristics that differed between elements included: Type of simulation, Data collection and analysis methodology, Number of pilots participating in on-line test runs, as well as Number and interactivity of simulator bridges. Advanced simulation technologies were employed, and tailored to the needs of each study element: Six-Degree of Freedom ship simulation math model; Frequency domain vertical motion analysis for a vessel in open shallow water, including the second order wave-induced "set-down"; and interactive simulator bridges. This paper reports the research methodology, findings, conclusions and recommendations for each study element; it also reports how the findings and recommendations have been adopted in the channel configuration. In addition to the conclusions relating to waterway design for each, the collective results illustrate the impact of the different research methodologies and simulation technologies on the type and quality of research findings. The paper addresses the simulator research issues associated with the advanced simulation technologies, and concludes with a list of challenges to simulator research.

　

 The US Army Corps of Engineers (ACOE), head-quartered in Washington, DC, has 8 Divisions all over the US with 38 district offices in the US, Asia and Europe. Division and Headquarter offices are involved with policy decision, national-level budgeting, legislative and executive branch coordination, etc. The District offices perform much of the technical engineering effort. The Norfolk District (ND) is one of six districts within the North Atlantic Division as illustrated in Figure 1. The ND performs many military and civil works functions within the state of Virginia. Included among the ND's civil works responsibilities is construction and maintenance of Congressionally authorized federal navigation channels servicing the Ports of Hampton Roads, Virginia.

　

 A local cost-sharing sponsor supports each of the District's navigation projects. The local sponsor for the federal navigation channels serving the Ports of Hampton Roads is the Virginia Port Authority (VPA). The VPA participates in a project by contributing funds, real estate, and other items of local cooperation for constructing navigation channels.

 The ND and the VPA recognize that Post-Panamax container vessels will soon call on Virginia's Ports. The Norfolk Harbor Federal Navigation Channel System allowing access to Virginia's ports requires improvement to provide service to these large ships.

　

　

 As illustrated in Figure 2, the current channel configuration is asymmetric, providing a 50'-element outbound (South Channel) and a 45'-element inbound (North Channel). For about 15 years this configuration has proven successful for navigation of large coal ships arriving in the port light-loaded, and departing the port fully laden. However, arrival of the next generation deeper-draft containerships meant visiting ships would require deeper channel depths on both the inbound and outbound transits.

　

　

 To determine the design requirements for the channel system, the ND office partnered with the Computer Aided Operations Research Facility (CAORF) at the US Merchant Marine Academy (USMMA) in Kings Point, NY, to perform ship simulation studies to support the channel deepening effort. The Corps' Engineer Research and Development Center (ERDC) at the Waterways Experiment Station provided technical data, oversight, and review functions for the study. The effort required complex coordination with the VPA, the Virginia Pilot Association, National Oceanographic and Atmospheric Administration (NOAA), U.S. Coast Guard (USCG), U.S. Navy, and other interests.

　

 The District's overall study objectives were to provide safe navigation channels, while looking for ways to reduce dredging costs. Due to time constraints associated with the overall project, the District had to prioritize the issues to be investigated. The District chose to investigate three primary areas: Thimble Shoal Channel (TSC), Atlantic Ocean Channel (AOC), and Norfolk Harbor Reach (NHR), as illustrated Figure 3.

　

 TSC was studied to validate the safety of two large vessels meeting in a confined channel. AOC was studied to determine under keel clearance (UKC) requirements due to the vertical motions induced on the vessel by wave and wind action, currents, vessel response, and other factors. NHR was studied to determine channel design requirements related to vessels meeting in turns and operating at relatively slow maneuvering speeds.

　

 The background, methodology, findings, conclusions, and recommendations for each of these elements are reviewed in the following sections. The details of these studies are reported in [1], [2], and [3].

　

　

2.1 Background

　

 The incremental deepening of the TSC in Chesapeake Bay is one element of the Norfolk port improvement program. Initial plans are for the inbound lane of this channel to be dredged to a depth of 50', matching the existing depth of the outbound lane. This would accommodate the needs of deep draft containerships calling on the port. TSC is a long stretch, about 11.7 miles in length. The area of major concern is near buoy pairs #11 and #12, as illustrated in Figure 4, where the areas outside of the channel are shallow. The purpose of this study was to evaluate the 50' deep configuration of TSC, when used by a large fully loaded inbound containership, meeting a large loaded outbound collier in this higher risk area.

　

　

2.2 Methodology

　

 This study used a man-in-the-loop simulation approach. It was conducted using the CAORF full-mission Visual Bridge Shiphandling Simulator (VBSS). The evaluation was based on comparing the performance during simulated transits of a fully loaded S-Class ship (as the new design vessel) in the proposed 50' deepened channel (as the new design channel); with that of a partially loaded S-Class ship that approximates the largest containership in existing vessel traffic (as the baseline vessel), in the existing TSC channel (as the baseline channel), with a minimum depth of 45' in the inbound lane and 50' in the outbound lane. The ownship models had Six Degrees of Freedom (6-DOF). The fully-loaded S-Class ship was configured to have a draft of 47'. During validation runs it became clear that the fully loaded S-Class ship would not be able to transit at the speeds wanted by the pilot, due to a high potential for grounding in the 50' channel. Thus the new design channel was tested with 52' (50' plus 2' of dredging allowance).

　

 The main environmental test conditions consisted of worst credible wind (50 knots) and current (1.5 knots) from the starboard beam, and abaft of the beam. Calm conditions were also simulated in some baseline transits.

　

 The safety criterion traditionally accepted in simulation studies is the comparative performance of the new design (new ship and waterway characteristics) versus the baseline design (existing ship and waterway characteristics). The baseline design is considered safe, by default, since it has performed acceptably in the past and continues without question today. This relative measure of safety, which is considered the most valid and typically available, enables data to be collected for comparative safety evaluation.

　

 One senior Virginia Pilot conned the inbound ships from the bridge of the simulator in all of the transits. The outbound traffic vessel in this study followed a predetermined worst credible scenario path, similar to the methodology used in numerous preceding CAORF studies. Note that the TSC study was conducted before the second bridge of CAORF was built. Questionnaire data were collected from the pilot before, during and after the series of simulated transits, to augment the simulator generated data.

　

 A series of 29 simulated transits were conducted, 8 in the baseline vessel, 20 in the new design vessel and one in a smaller auxiliary oiler. A total of 25 of these 29 runs were used for descriptive statistical analysis. The data analyzed included objective data from each simulated transit, and pilot judgments. The findings and conclusions of this study are based on the judgments of the Norfolk pilot in collaboration with the CAORF staff, and as interpreted by the staff.

　

2.3 Findings

　

 The evaluation was made in three parts: 1) channel navigation, 2) meeting and passing, 3) channel depth:

　

 Channel Navigation. The fully loaded S-Class ship was found to handle better than the partially loaded S-Class ship under the severe wind and current conditions. However, the two ships were judged about equivalent from a pilotage standpoint, since the partially loaded ship was able to transit at a higher speed and thus attain equivalence in controllability.

　

 Meeting and Passing. The pilot was able to adequately control the fully loaded S-Class ship, under the severe wind and current conditions, to successfully meet and pass the large outbound ship in multiple transits. Although some differences were noted between the fully loaded and partially loaded S-Class vessels, on balance the performances with both vessels were approximately equivalent. On this basis, the evidence suggests that the safety of meeting and passing a large outbound ship is equivalent for fully loaded and partially loaded S-Class ships under severe wind and current conditions.

　

 Although this investigation was not designed to address the width of the TSC, but rather to investigate the comparative performances of the fully loaded and partially loaded S-Class ships, some evidence was found to suggest that passing distances achievable between large ships in the TSC are less than the lower acceptable safety margin identified by this one pilot. This finding, which pertains to both the baseline conditions and the new design conditions, raises the issue of channel width.

　

 Channel Depth. Channel depth was not originally an issue of the investigation. However, during validation and screening runs it became apparent that the fully loaded 47' S-Class ship would not be able to transit at the higher speeds desired by the pilot, due to a potential for grounding in the 52' channel. The slim dynamic under keel clearance (DUKC) at 12 knots would often cause this large ship to ground on the simulator, due to roll associated with the severe wind and the large rudder control activities. DUKC refers to the UKC when the ship is underway or in motion. Even at 10 knots, the speed at which many of the test transits were made, the ship grounded after passing the traffic ship due to the roll associated with the ship-to-ship interaction. Subsequent analysis, however revealed that the ship model produced greater roll from the passing interaction than should occur at sea.

　

 The high propensity of grounding experienced in the simulation was likely exaggerated, due to insufficient roll damping or other limits of the ship simulation model in very shallow water. However, the very limited DUKC presents a real grounding risk in conditions that may cause substantial vertical or roll motions.

　

2.4 Conclusions

　

 With regard to channel navigation, the inbound transit of the fully loaded S-Class ship in the 52' deep TSC was concluded to be acceptably safe for wind under 50 knots, since its control was equivalent to that of the partially loaded ship in the existing channel, thus meeting the criterion of relative safety. The safety associated with the potential of this vessel grounding is an important, but separate, issue.

　

 At severe wind conditions (above 50 knots) the safety of an inbound transit in the fully loaded S-Class ship should be considered marginal, with caution exercised in making the decision to bring the ship in. This is due to a potential for grounding in a 52' deep channel, resulting from the very limited DUKC and the potential for the ship to roll/heave/ pitch, particularly when meeting another large ship.

　

 The results further suggest that transits of large ships of the S-Class size are marginal today, when meeting a large traffic ship under severe cross wind (above 50 knots) and current conditions. The width of the existing TSC was concluded to be adequate under most circumstances, but marginal for very large ships meeting under crosswinds above 50 knots.

　

2.5 Recommendations

　

　

2.6 Adopted Measures

　

 Based on the simulator study, the ND adopted the following measures to address the identified issues:

　

 Dredging plans will be to provide a 50' channel for over the 1,000' authorized width of the channel, not the 52' channel modeled in the Ship Simulation Study. A 52' channel would require dredging on both the inbound and outbound lanes since the existing channel only provides 50' outbound. Dredging both lanes was beyond the VPA's anticipated participation in the project and the VPA stated their participation in the project was limited to a 50' channel. After the VPA clarified the expectations for the project the draft of the design vessel was decreased to accommodate a shallower project depth.

　

 Construction contract dredging depths will require 51' plus one foot allowable overdepth, as is standard practice on most of the ND contracts. The construction depth was determined by examining the TSC's extensive dredging history, customer requirements, and other navigation issues.

　

 The design vessel's underkeel clearance requirements were documented in the ND economic evaluation report. Defense of the design vessel's underkeel clearance requirement was challenging for ND team members. The simulation study was an invaluable tool for providing ND channel designers with vessel depth requirements attributable to vessel roll, squat, salinity effects, and safety clearance. Without the simulation study the defense of UKC requirements would be severely weakened if not impossible.